OBJECTIVES Thermal ablation is an alternative treatment to cryotherapy within ‘screen and treat’ cervical screening services using visual inspection with acetic acid (VIA) in resource-constrained settings. This ablative treatment is suitable for low-grade squamous epithelial lesions. We evaluated the effectiveness of thermal ablation in the treatment of VIA-positive lesions within a ‘screen and treat’ programme in Malawi, and its acceptability to clients and providers.
METHODS Over the last several years, the Nkhoma Cervical Cancer Screening Programme has implemented a ‘screen and treat’ approach using VIA and treatment using thermal ablation in a rural district general hospital and associated health centres. Women with suitable VIA-positive lesions are offered treatment with thermal ablation; treated women return for review at one year. A patient experience questionnaire using validated facial pain scales was developed and translated into Chichewa.
RESULTS Recurrence rate (i.e. VIA positivity) was assessed in a cohort of over 580 treated women who returned for a 1-year review visit: among HIV negative women (n=546) 94.3% were VIA-negative, i.e. a treatment failure rate of approx. 5-6%, comparable with the international literature. Among women living with HIV (n=133), 91% were recurrence-free: in line with international findings and supporting the need for careful follow up. Staff reported professional satisfaction in being able to offer treatment consistently to VIA-positive clients, closer to their communities. Over 120 women have completed pain scales questionnaires following treatment with the traditional WISAP machine, or with one of the two new hand-held models (WISAP, or Liger).
CONCLUSIONS In many low-resource settings, VIA-based screening with robust treatment protocols will remain central to cervical cancer control until the promise of HPV vaccination and HPV testing are fully realised. Thermal ablation is an effective treatment modality, acceptable to clients and patients – however, quality assurance and provider mentoring are required for long-term effectiveness.

OBJECTIVE To determine the spectrum and time to occurrence of second primary cancers (SPC) stratified by HIV status.
METHODS Using linked cancer and HIV data from the South African National Cancer Registry and the National Health Laboratory service respectively; we identified individuals with first and second primary cancers. The sequence of cancer occurrence was determined from the cancer diagnosis date in patients with SPCs. We excluded individuals with cancers diagnosed on the same day (n=281). We compared median time to SPC diagnosis between HIV positive and negative individuals using the non-parametric Wilcoxon rank sum test. As a sensitivity analysis, we excluded non-melanoma skin cancers.
RESULTS Of the 292 589 public sector records, 7819 patients had more than one primary cancer. Excluding non-melanoma cancers, (n=3100), prostate 14% (n=663), breast 13% (n=602), colorectal 7.1% (n=330) and cervix 6.3% (n=294) cancers were the top SPCs in the cancer database. The HIV status was known for 1469 (19.5%) SPC patients, of which 449 (31%) were HIV positive. Amongst HIV positive people, cervix 14% (n=47), non-Hodgkin’s lymphoma 12% (n=38), breast 12% (n=38) and Kaposi sarcoma 11% (n=35) were the most common SPCs. For HIV negative individuals, breast 18% (n=115), prostate 11% (n=73), colorectal 7.7% (n=51) and cervix 6.7% (n=44) were the leading SPCs. Adjusting for first primary cancer diagnosis older age and white ethnicities compared to black ethnicities were predictors for SPC. Amongst those with SPCs, the median time to SPC diagnosis was 19 months (IQR: 5-46) among HIV positive individuals, 23 months (IQR: 9-48) in HIV negative individuals and 28 months (IQR: 9-57) for those with an unknown HIV status. The median time to SPC diagnosis was 9 months shorter in black ethnicities compared to white ethnicities (p<0.001).
CONCLUSIONS Cancer prevention and screening should remain a priority even in individuals previously diagnosed with cancer.

OBJECTIVE Women living with HIV (WLHIV) experience decreased survival from breast cancer. We sought to determine whether WLHIV surviving breast cancer also experienced decreased post-treatment quality of life (QOL).
METHODS We included women enrolled in Thabatse Cancer Cohort presenting from October 2010 to September 2018 at oncology centres in Botswana for initial treatment of breast cancer. QOL was measured quarterly using the SF-8, which includes a physical composite score (PCS) and a mental composite score (MCS). We assessed change in PCS and MCS from 3 months post-treatment to 18 months post-treatment. We performed multivariable linear regression to model the relationship between change in QOL and HIV, age, wealth, and cancer stage.
RESULTS Of 478 women enrolled, 134 women died prior to 18 months post-treatment and 125 women had not reached 18 months post-treatment. A total of 31 (14.2%) women missing QOL measurements were excluded. Of the remaining 188 analysed, 45 (23.9%) were WLHIV and 143 (76.1%) were HIV-uninfected. Almost half (48.1%) had advanced stage (stage III/IV) and 84.6% received multimodality treatment with surgery and chemotherapy and/or radiation. Overall, PCS increased following treatment, + 2.22 (95%CI 1.03 - 3.40), and MCS was largely unchanged, + 0.87 (95%CI -0.57 - 2.31). HIV was not associated with impaired recovery in PCS or MCS. A non-significant association between HIV infection and improvement in PCS was observed, +2.48 (95%CI -0.34 – 5.30, p=0.087). HIV was not predictive of change in MCS, -0.073 (95%CI -3.56 – 3.42, p=0.97). In the secondary analyses of the components of the PCS, HIV was predictive of improved general health (p=0.037), but not of the other four PCS components. Younger age and increased wealth were non-significantly associated with improved PCS recovery.
CONCLUSIONS Concurrent HIV infection does not appear to impair recovery in quality of life among women surviving breast cancer.

OBJECTIVE To determine the distribution and influence on survival of Human Papilloma Virus (HPV) types, lineages, and sub lineages in invasive cervical cancer in Uganda.
METHODS We recruited 264 patients with invasive cervical cancer (94% squamous cell carcinoma (SCC)) diagnosed at the Uganda Cancer Institute between July 2013 and February 2018. Tissue removed during biopsy was stored in liquid nitrogen until assay for HPV types (Roche Linear Array) and variants (PCR-based direct sequencing of the LCR/E6/E7 region of each type) at the University of Washington.
RESULTS Based on analysis of 241 subjects to date, we detected HPV DNA in 92% of cervical cancers. The distribution (not mutually exclusive) of HPV types was 42% HPV16, 22% HPV18, 11% HPV45, 4% HPV35, 3% HPV33, 3% HPV31; 62% contained either HPV16 (41%), HPV 18 (20%), or both (1%). Among the HPV16 positive tumours (n=101), the distribution of lineages was 4% A, 67% B, 24% C, and 4% D. Among the HPV18 positive tumours (n=51), the distribution of lineages was 12% A, 82% B, 2% C, and 4% D. In univariate analysis, there was a trend for HIV-infected patients to have lineage C compared to lineage B (p=0.06); otherwise no other clinical factors were associated with HPV type. Overall survival was not associated with HPV lineage (HR=1.3; 95% CI = 0.7, 2.9).
CONCLUSIONS Nearly two-thirds of cervical cancers (largely SCC) were associated with HPV16 and HPV18. These results are similar to series of SCC from other parts of the world. The distribution of HPV16 lineages is consistent with what has been previously observed from African-descent populations located in the U.S. HPV lineages were not associated with differences in clinical presentation or survival outcomes based on our preliminary analyses.

OBJECTIVE We are conducting a cervical cancer screening study of ~5,000 HIV+ women, aged 30-54 years, living in Rwanda to compare different screening and management strategies.
METHODS During the screening visit, a nurse administers a questionnaire on demographics and Invasive Cervical Cancer (ICC) risk factors, collects a specimen for HPV testing by GeneXpert and does Visual Inspection with Acetic acid (VIA). At colposcopy for screen-positive women, two additional specimens are taken for biomarker evaluations (E6/E7 oncoproteins) followed by rigorous colposcopic evaluation including 4-quadrant microbiopsies/biopsies. In a subset of women (n=720, 45% of screen-positive women) for whom histologic diagnosis is now available, the relative sensitivity for cervical intraepithelial neoplasia grade 2 or more severe diagnoses (CIN2+) was calculated for HPV DNA testing and VIA. The relative sensitivity and specificity for CIN2+ among HPV+ women for DNA and E6/E7 detection of HPV16, HPV1618/45 and HPV16/18/31/33/35/45/52/58 were calculated.
RESULTS The prevalence of HPV DNA was 26.7% and VIA positivity was 10.3%. The prevalence of CIN2+ was 6.4% (46 of 720) and ICC was 1.3% (9 of 720). Relative sensitivity of HPV DNA and VIA for CIN2+ was 89.1% and 43.5% respectively, p<0.001. Among HPV+ women, DNA and E6/E7 detection of HPV16 were 26.3% and 17.5% sensitive and 88.6% and 96.2% specific, respectively, for CIN2+. DNA and E6/E7 detection of HPV16/18/45 were 39.5% and 27.5% sensitive and 76.4% and 91.4% specific, respectively, for CIN2+. DNA and E6/E7 detection of HPV16/18/31/33/45/52/58 were 84.2% and 65% sensitive and 32.4% and 69.8% specific, respectively, for CIN2+.
CONCLUSIONS HPV DNA testing was twice as sensitive for CIN2+ but 2.5-fold more likely to be positive compared to VIA. Sensitivity increases and specificity decreases for CIN2+ with DNA and E6/E7 detection of more HPV types, with DNA being more sensitive but less specific. Future results will include more pathology results and the evaluation of other biomarkers.

BACKGROUND Coinfection with hepatitis B virus (HBV) and human immunodeficiency virus (HIV) is common in Sub-Saharan Africa (SSA) and can lead to progression of liver disease, cirrhosis and hepatocellular carcinoma (HCC). In Uganda HCC is the 3rd commonest cancer in males and 4th in women. 60-80% of HCC in Africa is attributed to HBV. Recent data demonstrate ongoing HBV transmission among HIV-infected adults in SSA, suggesting that HCC in the setting of HIV/HBV coinfection could be prevented with prior HBV vaccination. Because HBV vaccine efficacy is poorly understood among HIV-infected persons in SSA, we sought to characterize the humoral response of HIV-positive Ugandan adults to the vaccine.
METHODS We enrolled HIV-infected adults in Kampala, Uganda with no serologic evidence of prior exposure to HBV. Baseline socio-demographic variables, HBV and HIV history were obtained by participant interview and medical chart review. Three 20µg HBV vaccine doses were administered at intervals of 0, 1 and 6 months. Anti-HBs levels were measured at 4 weeks after the third vaccine dose. “Response” to the vaccine was defined as anti-HBs levels ≥10 IU/L and “high response” as >100 IU/L. Univariate and multivariate regression analysis were used to determine the predictors of vaccine response.
RESULTS Of 251 HIV-positive adults screened, 132 (53%) had no prior HBV-infection or vaccination and were enrolled. Most participants were women 89 (67%) and had a median (IQR) age of 32 years (27-41). 68 (52%) were on antiretroviral therapy >3 months. Median (IQR) CD4 count was 426 (261-583), and 64 (94%) of the 68 receiving ART had undetectable HIV RNA. Overall, 117 (92%) seroconverted to the vaccine (anti-HBs titers ≥10IU/L) with majority 109 (86%) being high-level responders (anti-HBs ≥100IU/L). Eight (6.3%) were low-level responders (anti-HBs 10-100IU/L) and 10 (7.9%) non-responders (anti-HBs <10). In univariate analysis, only baseline CD4 >200 cells/mm3 was associated with both response (p=0.02) and high-level response (p=0.01). In multivariate analysis baseline CD4 was associated with “high response”, with each unit increase in CD4 count corresponding to a 0.004 increase in vaccine response (p=0.01, CI 0.001-0.008).
CONCLUSIONS Half of the screened patients did not have immunity to HBV-infection, suggesting a large “at risk” population for the infection among HIV-positive adults in Uganda. The vaccine was effective in eliciting a protective humoral immunologic response, particularly among those with higher CD4 counts. Our findings support routine HBV vaccination among HIV-positive adults at risk for HBV infection in a bid to prevent HCC.

OBJECTIVE Kaposi sarcoma (KS) is an AIDS-defining cancer and a cause of enormous cancer morbidity and mortality in Sub-Saharan Africa. KS lesions comprise a complex tumour ecosystem with heterogeneous malignant and non-malignant cellular components. Characterizing the elements that comprise KS tumours and how these components interact in vivo will advance our understanding of KS tumorigenesis.
METHODS We evaluated KS tumour and normal skin samples obtained from HIV-positive and HIV-negative adults with KS enrolled in an ongoing study at the Uganda Cancer Institute – Fred Hutch Cancer Centre in Kampala, Uganda. We have developed and implemented a platform for isolating and interrogating viable single cells from KS biopsies. Cryopreserved single cell suspensions are subsequently analyzed on a BD FACSAria™ II sorter, using a panel of antibodies to identify non-malignant cells such as CD3+CD8+ T-cells and candidate KS tumour cells carrying the CD34+/VEGFR3+/LYVE-1+ surface phenotype. To date, targeted transcriptional profiling and T-cell receptor-alpha (TRA) / T-cell receptor-beta (TRB) variable region sequencing was performed on single CD8+ T-cells from KS tumour biopsies.
RESULTS Transcriptional profiling of single CD8+ T-cells using multiplex RT-PCR with primers for 24 genes relevant to lineage, function, proliferation, or exhaustion, followed by high-throughput sequencing of indexed PCR products, revealed significant heterogeneity in the expression of many genes, but uniformly low expression of several genes associated with proliferation and functional activation, including Ki-67, granzyme B, and TNF-alpha. Sequencing of TRA and TRB CDR3 regions carried in single CD8+ T cells from KS biopsies revealed several TRB sequences that our group has tentatively identified as candidate “public” KS-specific TCRs.
CONCLUSIONS Our preliminary data suggest that the transcriptional profile of CD8+ cells in KS tumours is consistent with an “exhausted” profile, which may have implications for the use of therapeutic agents for reversing T-cell exhaustion in the treatment of KS.

OBJECTIVE HPV self-sampling is widely recognized as a sensitive test to detect precancerous lesions and well-accepted approach to cervical cancer screening. Self-sampling is particularly well-suited to low-resource settings as it can help achieve higher screening coverage. Our objective is to describe self-sampling uptake within an effort to deliver HPV screening in public health systems in Nicaragua, Guatemala, and Honduras.
METHODS PATH and country nongovernmental organizations worked with the ministries of health to implement HPV-based screening for cervical cancer. Self-sampling was implemented using QIAGEN’s careBrush. Key screening program indicators were compiled, including: number of women screened, number who used self-sampling, women’s age at screening, and time elapsed since the woman’s last screening test. Data were consolidated across years 2015 to 2018 and countries.
RESULTS During the project, 234,078 women across the three countries were screened with HPV tests; of these, 147,703 (63.1%) used self-sampling. HPV test positivity was 13.6%, with no differences in overall positivity between self-sampling and clinician-collected samples. In Nicaragua, 95.8% of women screened used self-sampling and 47.3% of women reached had never been previously screened. In Guatemala, where data are available, 90.2% of women used self-sampling and 30.0% were screened for the first time. In Honduras, self-sampling was promoted in just one of the three target provinces starting in 2017; in that province from 2017–2018, 58.4% used self-sampling.
CONCLUSIONS Self-sampling was highly accepted and reached many never-screened women, particularly in Nicaragua where health system leaders strongly encouraged self-sampling. Self-sampling may present a unique opportunity to reach high screening coverage in low-resource settings using HPV-based screening approaches.

OBJECTIVE Incidence of squamous cell carcinoma of the head and neck (HNSCC) in Botswana has more than doubled over the past decade. We sought to assess association between HIV and other established HNSCC risk factors and the risk of developing HNSCC in Botswana.
METHODS As part of the Thabatse Cancer Cohort (TCC), we included patients presenting with HNSCC from 2010-2019 at four principal oncology treatments centers. Age (in 5-year categories) and sex matched controls (matched 4:1) were drawn from a 30-community random population sample including 12,600 rural and peri-urban residents. Conditional multivariable logistic regression was used to assess impact of HIV, smoking, and use of indoor solid fuels on HNSCC risk. Models adjusted for income and age (as continuous) to reduce risk of residual confounding.
RESULTS A total of 191 HNSCC cases were enrolled, including 166 (87%) males and 25 (13%) females. The median age was 59 years (IQR 49-67). The majority of cases (147, 79%) reported having ever smoked, and 67 (37%) of cases were HIV-infected at time of cancer diagnosis. In adjusted analysis, smoking was strongly associated with HNSCC risk, aOR 8.64 (95% CI 5.0 to 14.9 p<0.001). HIV infection was also significantly associated with risk of HNSCC, aOR 2.06 (95% CI 1.2 to 3.6). Current use of solid fuels was associated with increased HNSCC risk, aOR 1.74 (95% CI 0.99 to 3.1), but this finding did not reach statistical significance (p=0.055). Smoking is attributable for 71% of HNSCCs in Botswana.
CONCLUSIONS Smoking and HIV are important risk factors for HNSCC in Botswana, and the increasing use of tobacco, particularly among persons living with HIV, may account for the rise in HNSCC incidence. Exposure to indoor solid fuel use may also increase HNSCC risk, but the limited number of cases prevents firm conclusion.

BACKGROUND High-risk human papilloma viruses (hr-HPV) are the primary cause of cervical cancer. HPV vaccination is expected to prevent cervical cancers caused by HPV types included in the vaccines, and perhaps by cross-protection from other types. This study was designed to determine the hrHPV type distribution in women at two rural communities in Zimbabwe.
METHODS We implemented a cross-sectional study in two rural communities in Zimbabwe. 618 community-based self-collected swabs were collected in the Chidamoyo region. Samples were initially analyzed by Cepheid GeneXpert with hrHPV typing analysis performed using SeeGene Anyplex RT-PCR. At the Karanda Mission Hospital, clinician-collected cervical swabs were obtained from 400 women presenting for visual inspection with acetic acid screening; hrHPV analysis was as described for the Chidamoyo cohort.
RESULTS Chidamoyo
• hr-HPV prevalence: 17% (105/618); 34% (39/115) vs. 13% (66/503) among HIV-1-positive vs. -negative participants, respectively
• HPV 35 was detected most frequently (20%); an equal percentage of HPV16, HPV18, HPV52, and HPV 68 were detected (16%).
• 30% of women had an infection with HPV16 and/or -18 (included in the bivalent vaccine)
• Only 40% of the women had an infection(s) that would be included in the 9-valent vaccine types
Karanda
• hrHPV prevalence: 17% (67/400); 28% (19/68) vs. 17% (69/400) among HIV-1 positive vs. negative participants, respectively
• 8/17 (48%) hrHPV isolates from HIV+ women not covered by vaccines in current use
• HPV16 -most frequent type 22% (15/67); second HPV35 - 18% (12/67) and HPV 18 - 10% (7/67)
• 45% (30/67) harbored an hrHPV type not covered by the vaccines
CONCLUSIONS As many as 45%-60% of hrHPV isolates from women in both rural hospitals are types not covered by the bivalent (HPV16/18) or 9-valent vaccines. HPV 35 was detected in 18 and 20% of isolates at Karanda and Chidamoyo respectively, and HPV 16 was detected in 22 and 16%, respectively.